The article examines: (i) the reasons of error due to thermoelectric inhomogeneity of electrodes of thermocouples acquired during prolonged use; (ii) the neural network method of error correction based on a generalization of verification results in several temperature fields; (iii) the method of investigating the impact of changing the speed of the conversion characteristic drift of thermocouple on error correction; (iv) results of this investigation. It is shown that residual error for type K thermocouples at the 5 % level of significance does not exceed μ±0.46 oС and one at the 10 % level of significance does not exceed ±0.25 °С
The simplified metrological software test (MST) for modeling the method of determining the thermocouple (TC) error in situ during operation is considered in the paper. The interaction between the proposed MST and a temperature measuring system is also reflected in order to study the error of determining the TC error in situ during operation. The modelling studies of the random error influence of the temperature measuring system, as well as interference magnitude (both the common and normal mode noises) on the error of determining the TC error in situ during operation using the proposed MST, have been carried out. The noise and interference of the order of 5-6 μV cause the error of about 0.2-0.3°C. It is shown that high noise immunity is essential for accurate temperature measurements using TCs.
Error due to inhomogeneity is the main problem of thermocouples (TCs), e.g., during the operation of a type K TC, this error can reach 11-30 °C. Thus, metrological reliability of TCs is threatened by this error because there is a high risk of exceeding the permissible error when the temperature distribution along the TC legs changes. Such a large error, in turn, can threaten a proper operation or even safety of a measured object. A TC with controlled temperature field was proposed to cope with this error. An information-measuring system to perform proper measurements, measurement data acquisition and collection to construct mathematical models is proposed. Its property is high diurnal stability of ±(0.0025+0,002(X/XMAX–1) %. The requirements for the information-measuring system and its structure are considered in this paper. In particular, one of the key problems of such a sensor is how stable is its own temperature field under the influence of the temperature field of a measured object. The experimental studies were carried out using the developed system. They showed that the coefficient of penetration of the temperature field of the measured object is about 0.04. This allows decreasing error due to inhomogeneity by about 10-20 times.
The article presents the modified structure of the two-level digital frequency synthesizer (TLDFS), which combines the properties of classical digital frequency synthesizers (DFS) and Poisson pulse sequence generators (PPSG). The analysis of the statistical characteristics of synthesizer output signal, obtained in computer modelling with the use of appropriate software, has been carried out, which allowed determining the effective range of values of its control codes. The proposed generators can be effectively used to simulate various natural and technical processes, in particular, to simulate the output signals of dosimetric detectors during the design, adjustment and testing of dosimetric devices.